Ó Springer 2006

Mycopathologia (2006) 161: 385–394 DOI 10.1007/s11046-006-0020-2

Allee effect in the infection dynamics of the entomopathogenic fungus Beauveria bassiana (Bals) Vuill. on the beetle, Mylabris pustulata K. Uma Devi & C. Uma Maheswara Rao Department of Botany, Andhra University, Visakhapatnam, 530 003 AP, India Received 16 August 2005; accepted in revised form 24 January 2006

Abstract Successful infection by Beauveria bassiana as with all other entomopathogenic fungi, is accomplished only at a high conidial dose while, theoretically, a single conidium should be sufficient. Indeed, this is a major deterrent in its use as a biocontrol agent. High pathogen load for infection is required by organisms which display ‘Allee’ effect. In such organisms, a threshold exists for pathogen dose, below which no infection can be caused. B. bassiana has a semelparous life cycle and, therefore, its infection dynamics are expected to conform to the mass action principle with a linear relationship between dose and successful infection observable as mortality of the insect. Whether the need for a high conidial dose to induce insect mortality by B. bassiana is due to the operation of Allee effect was examined. A sample of 34 isolates was bioassayed on Mylabris pustulata (Coleoptera: Meloidae) at four conidial concentrations. With more than half of the isolates in the sample, the lowest dose tested (104 conidia/insect) did not cause insect mortality. Thus, a threshold pathogen load is required to cause successful infection. In these isolates, the dose–mortality relationship was sigmoid. Allee effect is thus identified in the infection dynamics of B. bassiana–M. pustulata system. The isolates that induced mortality at the lowest dose tested are concluded to be highly virulent with a lower threshold dose required for successful infection. With some isolates, at high conidial dose, the infection rate decreased either due to a decrease in the proportion of insects showing mycosis, to the speed of death, or both. Such a response could result from intra scramble competition arising from overload of pathogen at very high dose. Key words: high effective conidial dose, semelparous life history, sigmoid dose–mortality response, threshold dose

Introduction Beauveria bassiana, an ubiquitous insect pathogenic fungus with a very wide host range is marketed as a biopesticide. From a perusal of the several published reports of bioassays studying virulence of this fungus on insects [1, 2], and results on bioassays on nine insect species in our laboratory (unpublished), it has been observed that even under the most favorable conditions for infection, mortality is caused only when large

conidial doses are applied, even though a single conidium should theoretically be sufficient to induce infection. With most hyphomycetous entomogenous fungi, the regression slopes computed from probit analysis of dose–mortality data are low [3]. Consequently, a very high dose of 1013–1014 conidia of B. bassiana per hectare is used in field applications for effective knockdown of insect population [1]. Beauveria bassiana is an obligate killer. It has a semelparous life history with a single reproductive

386 episode. It releases all its infective propagules (conidia) in a single spell – after the death of the insect. The infection process of parasites with this type of life history is transparent – every successful infection results in host mortality and therefore can be detected. The infection dynamics of a semelparous parasite with clustering of all its (infective) propagules should conform to the ‘mass action principle’ [4], wherein infection rate is a linear function of the density of parasites and susceptible hosts. On the contrary, parasites with an iteroparous life history which have repeated spells of reproduction and a serial transmission of their infective units, can initiate infection only with a high dose of infective propagules. Such parasites display a so-called ‘Allee effect’ [5], a phenomenon that is characterized by an invasion threshold for the parasite, i.e., the parasite population can establish an infection only if its founder population size exceeds the invasion threshold [5]. In such systems that exhibit an Allee effect, the dose– infection relationship resolves into a sigmoid curve [5] in contrast to the linearity with systems in which infection dynamics follow the mass action principle. The Allee effect has been observed in infections of tree cutting ant by Metarhizium anisopliae, an entomogenous fungus closely related to Beauveria bassiana [4]. We tested whether the need for a large conidial dose to initiate infection by B. bassiana is also due to the existence of an Allee effect. To this end, bioassays were done on blister beetle (Mylabris pustulata Thunberg) (Coleoptera: Meloidae).

Materials and methods A sample of 34 isolates of B. bassiana (Table 1) was bioassayed against the blister beetle. Isolates from diverse insect hosts and geographic (climatic) regions were selected to constitute the sample to have a representation of the wide diversity within the species. This sample has been characterized in our laboratory for several other phenotypes and also molecular genetically characterized to assess the reproductive biology and phylogenetic affiliation of this species [6–10, Uma Devi et al., unpublished]. The fungal cultures were initiated on Sabouraud dextrose yeast agar (SDAY) slants from conidia stored in 20% glycerol at )20 °C. The conidia from 14-day-old cultures were used in

the bioassays. The conidia were harvested from the cultures by scraping with a sterile spatula, and 5 ml of an aqueous suspension with 0.02% Tween 80 (Sigma–Aldrich, India) was made by vortexing. An aliquot of the suspension was taken to check the viability of the conidia as described by Varela and Morales [11]. The remaining suspension was stored in a refrigerator (4 °C) for the treatments the next day. In a few isolates, the viability was also tested after storing in the refrigerator. No significant variation in viability was observed in the aqueous conidial suspension tested before and after storing at 4 °C. Cultures with more than 90% viable conidia were used in the treatments. The number of conidia per ml of the suspension was estimated through haemocytometer counts. Conidial suspensions of concentration of 105–108 conidia/ml were made. Adult beetles (M. pustulata) were collected from a heavily infested pigeon pea field using an insect trap. Care was taken to ensure that all insects collected were of comparable size to avoid large variations in age. Beetles were placed as a batch of 10 (for each treatment) in perforated plastic boxes (10  5 cm) with lids. Fresh pigeon pea flowers were provided as feed every day to the beetles. The boxes were cleaned of insect litter daily. They were placed in an environmental chamber set at 25 ± 1 °C, 90% relative humidity and an 8 h/16 h light/dark cycle. The insects were treated 2 days after collection. The beetles that were fatally injured during capture could be identified and removed in the 2 days before treatment. For treatment, the beetles were anaesthetized by exposing them to chloroform fumes from a cotton plug dipped in chloroform. The insects were treated singly with 100 ll of inoculum. The inoculum was dispensed with a micropipette (GilsonÒ) on the ventral side of the insect all over its body and head [12]. Treated beetles were put in boxes. Mortality was recorded daily for 30 days – the time up to which more than 90% of the control beetles were alive. There was no significant mortality in the treated insects after 14 days. Therefore, data up to 14 days after treatment was considered for analysis. The boxes with the treated insects were arranged in an environmental chamber set to conditions described above. The bioassays were set up as a completely randomized block design [13] with two replicates for each treatment. The

387 Table 1. The original host and geographic origin of the isolates of the sample of the entomopathogenic fungus Beauveria bassiana used in the bioassay on Mylabris pustulata Isolatea

Original host

Geographic origin

ARSEF 326 ARSEF 501 ARSEF 652 ARSEF 739 ARSEF 1038 ARSEF 1098 ARSEF 1149 & 1166 ARSEF 1169 ARSEF 1314, 1315 & 1316 ARSEF 1491 ARSEF 1512 ARSEF 1788 ARSEF 2860 ARSEF 3041 ARSEF 3120 ARSEF 3286 ARSEF 3387 NRRL 3108 NRRL 20698 NRRL 20699 NRRL 22864 NRRL 22865 NRRL 22866 ITCC 913 ITCC 1253 ITCC 4521 ITCC 4644 ITCC 4688 BB2 BB3 BB4

Chilo plejadellus Ostrinia nubilalis Ostrinia nubilalis Diabrotica paranoense Ostrinia nubilalis Nephotettix cincticeps Helicoverpa armigera Sitona lineatus Helicoverpa virescens Diatraea saccharalis Spodoptera littoralis Helicoverpa virescens Schizaphis graminum Reticulitermus flavipes Senecio sp. Spodoptera littoralis Myzus persicae Ostrinia nubilalis Dysdercus koengii Unknown Glichrochilus quadrisignatus Unknown Pachnaeus litus Unknown Musca domestica Diatraea saccharalis Oil palm larva Helicoverpa armigera Spodoptera litura Soil Helicoverpa armigera

Queensland, Australia China Beijing, China Goiania, Brazil New York, USA Japan Cordoba, Spain Senneville, France La Minie´re, France Lucighan, France La Minie´re, France Spain Idaho, USA Toronto, Canada Yvelines, France Montpellier, France Washington, USA Unknown Lima, Peru Illinois, USA Illinois, USA Iowa, USA Florida, USA The Netherlands Mumbai, Central India Karnal, North India Ambajipeta, South India Hyderabad, South India Bangalore, South India Bangalore, South India Warangal, South India

BB Isolates are from local (south Indian) fields and are not yet accessioned.

controls were treated with an equal volume of water with 0.02% Tween 80. The dead insects were transferred to moist chambers (autoclaved Petri dishes with a moist filter paper lining) to facilitate mycosis. The number of insects that expressed mycosis was noted. The bioassays were repeated once. The cumulative insect mortality in each treatment was corrected for control mortality [14]. The number of insects with mycosis was estimated as percent proportion of dead insects. The mortality and mycosis values were arcsine percent square root transformed to normalize the data [15]. The mean and standard error (S.E.) of all (four) replicates for mortality and mycosis in each treatment were calculated and arcsine back transformed. Median lethal time (LT50) was

calculated from the cumulative mortality data on each day post treatment, using survival analysis with Weibull distribution [16]. The percent mortality, percent mycosis and LT50 of each isolate at the four conidial doses were subjected to principal component analysis (PCA), a data reduction statistical method, to compute relative virulence index (RVI) of the isolate at each dose. The RVI value would indicate the infection rate of the isolate having been derived from all the three virulence parameters. In results which indicated the prevalence of an Allee effect, the dose (log value) mortality data was plotted as a graph to test for fit for a sigmoid curve. Statistical analysis was done using the computer adaptive statistical software packages – SPSS ver. 7.5 and STATISTICA ver. 5.0 [17, 18].

388

With 16 of the 34 isolates tested, no mortality was observed at the lowest (105 conidia/ml; 104 conidia/ insect) dose tested (Table 2). Thus, a concentration of 104 conidia per insect was not sufficient to induce successful infection. The mortality caused by five of these 16 isolates leveled off at values less than the theoretical maximum of 100% at the two high doses (107 and 108 conidia/ml). In the remaining isolates, 100% mortality was attained at the dose of 107 conidia/ml and remained the same at the higher dose (108 conidia/ml) (Table 2). Since no mortality was caused with these isolates at a dose of 105 conidia/ml, mortality is also not expected at a still lower dose (though this assumption was not tested). The dose–mortality relationship in these isolates thus resolved into a sigmoid curve (Figure 1). With the other 18 isolates, mortality was caused at the lowest conidial dose tested (105 conidia/ml), and it increased with increasing dose up to a dose of 107 conidia/ml, and beyond this dose, it leveled off in 14 isolates, in 11 among them, a theoretical maximum (100%) was reached by 107 conidia/ml; and it decreased in three isolates (Table 3). The isolates in which less than 15% mortality was caused at the lowest dose were also taken as suggestive of sigmoid response (Table 2). Dose–mycosis relationship One isolate, ARSEF 1038, was found to be nonmycotic at all the conidial concentrations tested (Table 1). With 14 isolates, no mycosis was induced in the dead insects at the low doses, 105 or 106 conidia/ml. With the other isolates, a positive correlation between mycosis and dose was observed up to a dose of 107 conidia/ml. A further increase in dose had a positive (18 isolates), negative (7 isolates) or neutral effect (8 isolates – in two of them 100% mycosis was produced at a dose of 107 conidia/ml) on mycosis (Tables 2 and 3). Relationship between dose and speed of kill The speed with which the insects were killed as indicated by LT50, increased with increasing dose

ARSEF 326 3

2

Probit Mortality

Dose–mortality relationship

up to a concentration of 107 conidia/ml in all isolates. At a conidial concentration of 108 conidia/ml, the speed of kill further increased (seven isolates), decreased (18 isolates) or leveled off (8 isolates) (Tables 2 and 3). No consistent relationship between speed of kill and mycosis was observed.

1

0

–1 –2

3

4

5

6

7

8

9

Log dose (spores per insect)

ARSEF 3286 1.5 1.0

Probit Mortality

Results

.5 0.0 –.5

–1.0 –1.5 –2.0 3

4

5

6

7

8

9

Log dose (spores per insect) Figure 1. The sigmoid dose–mortality relationship in Beauveria bassiana–Mylabris pustulata system in isolates showing an ‘Allee’ effect (a) isolate ARSEF 326 as a representative of isolates in which leveling off of insect mortality beyond a conidial dose occurred because the theoretical maximum of 100% is reached (y = )8.472 + 1.5281X + 0.188, r2 = 0.8829); (b) isolate ARSEF 3286 as a representative of isolates in which leveling off of mortality values beyond a conidial dose occurred even when theoretical maximum of 100% is not reached (y = )0.851 + 0.799X + 0.56, r2 = 0.8821). The graphs have been derived from probit-derived log values of dose and mortality (uncorrected for control mortality) using SPSS ver. 7.5 [17].

389 Table 2. Laboratory bioassay data demonstrating an Allee effect in the infection dynamics of the entomopathogenic fungus Beauveria bassiana on Mylabris pustulata Isolatea

Concb

108 107 106 105 ARSEF 652, NRRL22866, ITCC 4521 108 107 106 105 ARSEF 739 108 107 106 105 ARSEF 1038 108 107 106 105 ARSEF 1098 108 107 106 105 ARSEF 1315 108 107 106 105 ARSEF 1512 108 107 106 105 ARSEF 1788, ARSEF 3286, NRRL 3108, NRRL 20698 108 107 106 105 ARSEF 3387, ITCC 4688 108 107 106 105 NRRL 22864 108 107 106 105 ITCC 1253, BB3 108 107 106 105 ARSEF 326, NRRL 22865

a

Mor

Myc

LT50c

Rid

Response patterned

99.2 ± 0.8 99.2 ± 0.8 85.7 ± 0.4 0 62.5 ± 0.3 62.5 ± 1.5 52.2 ± 0.3 0 96.2 ± 3.8 98.3 ± 1.7 81.1 ± 1.6 0 37.6 ± 0.2 37.5 ± 0.1 37.8 ± 0.3 0 100 100 90.0 ± 0.4 0 56.3 ± 0.4 56.3 ± 3.4 48.6 ± 0.1 0 99.2 ± 0.8 99.2 ± 0.8 85.7 ± 0.4 12.5 ± 0.3 100 100 90.0 ± 0.4 0 99.2 ± 0.8 99.2 ± 0.8 85.7 ± 0.4 0 99.2 ± 0.8 99.2 ± 0.8 85.7 ± 0.4 0 100 100 48.6 ± 0.1 12.5 ± 0.3

50.0 ± 0.4 40.0 ± 1.6 0 0 40.0 ± 0.4 37.5 ± 0.5 33.0 ± 0.3 0 45.0 ± 0.5 44.0 ± 1.2 43.0 ± 0.3 0 0 0 0 0 100 100 80.0 ± 0.5 0 44.0 ± 0.3 42.5 ± 0.5 40.0 ± 0.2 0 40.0 ± 0.0 63.0 ± 2.6 38.0 ± 0.3 0 40.0 ± 0.4 33.0 ± 2.4 33.0 ± 0.3 0 10.0 ± 0.0 0 0 0 80.0 ± 0.4 50.0 ± 0.4 89.0 ± 0.1 0 100 40.3 ± 1.6 0 0

5.5 (4.5–6.3) 8.9 (6.8–10.2) 9.2 (8.6–11.3) e 13.1 (10.6–15.9) 11.2 (9.8–2.5) 11.5 (10.3–2.9) e 4.1 (3.1–4.9) 2.7 (2.3–4.5) 9.5 (8.5–10.6) e e e e e 6.6 (5.7–7.3) 6.2 (4.9–7.3) 9.9 (8.5–11.3) e 9.5 (7.7–11.9) 8.3 (6.2–9.3) e e 6.6 (5.7–7.2) 7.3 (6.1–8.5) 9.5 (7.5–11.3) e 4.2 (3.2–4.9) 3.4 (2.4–3.4) 9.8 (7.5–11.2) e 8.9 (7.6–9.7) 8.6 (7.6–9.7) 10.4 (8.6–12.1) e 6.3 (5.1–7.3) 3.2 (2.2–4.3) 5.5 (3.9–7.2) e 5.3 (4.4–6.1) 9.8 (4.9–11.0) e e

1.33 0.79 0.08 )1.14 0.46 0.89 0.61 )1.28 0.86 0.96 0.34 )1.25 )0.73 )0.73 )0.73 )1.33 1.04 1.08 0.57 )0.87 0.81 1.13 0.24 )1.21 0.79 1.02 0.34 )1.25 0.95 0.87 0.34 )1.25 1.29 0.66 0.21 )1.19 0.63 0.62 0.66 )1.29 1.70 0.74 )0.46 )0.76

RVI: Inc Mor: Sig Myc: Inc LT50: Dec RVI: Dec Mor: Sig Myc: Inc LT50: Inc RVI: Dec Mor: Sig Myc: Lev LT50: Inc RVI: lev Mor: Sig Myc: – LT50: – RVI: Dec Mor: Sig Myc: Levs LT50: lev RVI: Dec Mor: Sig Myc: Lev LT50: Inc RVI: Dec Mor: Sig Myc: Dec LT50: Dec RVI: Inc Mor: Sig Myc: Inc LT50: Inc RVI: Inc Mor: Sig Myc: Inc LT50: Lev RVI: lev Mor: Sig Myc: Inc LT50: Inc RVI: Inc Mor: Sig Myc: Inc LT50: Dec

Isolates with similar response pattern are clubbed. The values represent the actual values of the first isolate in the group. Conidia/ml: 100 ll/insect. c Values in brackets indicate fiducial limits, e = error, LT50 could not be computed because the mortality caused did not reach 50%. d Relative virulence index of an isolate at the four doses computed from mortality, mycosis and LT50 using principal component analysis, a data reduction statistical method. The RVI reflects the infection rate of the isolate. e Response pattern over the doses: Inc – increasing with increase in conidial dose; Dec – increasing with dose up to a concentration of 107 conidia/ml and decreasing at 108 conidia/ml; Lev – increasing with dose up to 107 conidia/ml with no further increase at 108 conidia/ml; Levs – increasing with dose and reaching the theoretical maximum (100%) at a concentration (usually) of 107 conidia/ml with no scope for increase with increase in conidial dose; Sig – sigmoid, with no mortality at the lowest dose, increasing mortality with increase in dose at the next two higher doses which finally levels off even with an additional increase in dose. For LT50, values with overlapping fiducial limits are taken as not significantly different. b

100 100 90.0 ± 0.4 25.0 ± 0.5 84.7 ± 0.3 84.7 ± 0.8 67.0 ± 0.1 50.0 ± 0.5 100 100 90.0 ± 0.4 25.0 ± 0.3 50.0 ± 0.5 75.0 ± 1.3 45.0 ± 0.3 0 59.4 ± 0.1 59.4 ± 0.1 50.4 ± 0.4 37.5 ± 0.1 100 100 90.0 ± 0.4 37.5 ± 0.5 60.6 ± 1.1 78.4 ± 1.2 90.0 ± 0.0 40.0 ± 0.5 96.2 ± 3.8 98.3 ± 0.7 81.1 ± 1.6 25.0 ± 0.3 98.3 ± 0.7 98.3 ± 0.7 83.9 ± 0.8 25.0 ± 0.3 100 100 90.0 ± 0.4 25.0 ± 0.3

108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105

ARSEF 501

ARSEF 3120

ARSEF 3041

ARSEF 2860, NRRL 20699

ARSEF 1491

ARSEF 1316

ARSEF 1314

ARSEF 1169

ARSEF 1166

ARSEF 1149

Mor

Concb

Isolatea

100 100 40.0 ± 0.4 50.0 ± 0.4 89.0 ± 0.4 100 78.0 ± 0.5 33.0 ± 0.3 20.0 ± 0.5 20.0 ± 1.0 0 0 50.0 ± 0.5 49.0 ± 1.3 60.0 ± 0.4 0 43.0 ± 0.0 49.0 ± 1.3 57.0 ± 0.3 0 10.0 ± 0.0 33.0 ± 2.4 33.0 ± 0.3 20.0 ± 0.0 25.0 ± 0.3 70.0 ± 0.5 50.0 ± 0.0 0 90.0 ± 0.4 70.0 ± 0.5 44.0 ± 0.3 20.0 ± 0.5 80.0 ± 0.4 89.0 ± 1.6 33.0 ± 0.3 50.0 ± 0.5 30.0 ± 0.5 13.0 ± 0.5 0 0

Myc

5.9 (5.1–6.6) 6.2 (3.9–7.3) 4.9 (3.9–6.2) e 3.6 (2.4–4.6) 5.1 (3.9–6.5) 5.1 (4.2–9.4) 11.3 (10.1– 14.7) 5.8 (4.9–6.4) 2.9 (1.5–3.8) 10.9 (8.5–15.6) e 10.4 (8.7–12.3) 7.9 (5.4–9.3) e e 9.5 (7.7–11.9) 7.6 (6.4–9.2) 6.5 (4.3–8.6) e 4.1 (3.0–4.7) 2.4 (1.5–3.6) 6.7 (5.4–8.2) e 8.7 (7.5–10.1) 8.4 (7.2–9.8) 8.5 (4.8–10.2) e 5.8 (4.7–6.6) 3.2 (2.1–4.4) 7.9 (5.2–9.2) e 6.8 (5.6–7.7) 6.7 (5.4–7.9) 8.7 (5.9–10.2) e 9.8 (8.7–10.7) 5.6 (4.5–6.6) 9.9 (8.5–11.2) e

LT50c

0.92 0.90 0.34 )1.25 1.10 1.08 0.49 )0.89 1.12 1.33 )0.02 )0.81 0.80 1.40 0.39 )0.86 0.88 0.92 0.35 )1.25 0.36 1.08 0.53 )1.25 0.02 0.86 0.79 )1.23 1.04 1.02 0.16 )1.19 1.03 1.15 0.06 )1.15 1.01 1.11 0.11 )1.17

RVId

RVI: Inc Mor: LevS Myc: LevS LT50: lev RVI: Inc Mor: Lev Myc: Dec LT50: Dec RVI: Dec Mor: LevS Myc: Lev LT50: Inc RVI: Dec Mor: Dec Myc: Lev LT50: Inc RVI: Dec Mor: Lev Myc: Dec LT50: Inc RVI: Dec Mor: Levs Myc: Dec LT50: Inc RVI: Dec Mor: Dec Myc: Dec LT50: Lev RVI: Inc Mor: LevS Myc: Inc LT50: Inc RVI: Dec Mor: LevS Myc: Dec LT50: Lev RVI: Dec Mor: LevS Myc: Inc LT50: Inc

Response patterne

Table 3. Laboratory bioassay data of isolates of the entomopathogenic fungus Beauveria bassiana on Mylabris pustulata in which an Allee effect was not observable*

390

108 107 106 105 108 107 106 105 108 107 106 105 108 107 106 105 50.0 ± 0.0 75.0 ± 0.5 90.0 ± 0.5 12.5 ± 0.3 50.0 ± 0.5 75.0 ± 0.3 45.0 ± 0.3 12.5 ± 0.3 100 100 45.0 ± 0.3 25.0 ± 0.0 100 100 90.0 ± 0.0 37.5 ± 0.3

87.0 ± 0.3 66.0 ± 0.6 80.0 ± 0.4 33.0 ± 0.3 50.0 ± 0.4 50.0 ± 0.4 100 33.0 ± 0.3 85.0 ± 0.3 75.0 ± 1.2 50.0 ± 0.5 90.0 ± 0.0 60.0 ± 0.0 86.4 ± 2.9 40.0 ± 0.4 44.0 ± 0.3

8.3 (6.5–10.5) 3.8 (2.8–4.9) 5.6 (4.5–7.8) e 7.5 (5.7–9.4) 5.6 (4.3–6.1) e e 5.9 (4.5–7.0) 5.5 (4.3–6.9) e e 6.4 (4.9–7.0) 8.8 (7.5–10.1) 9 (7.1–11.2) e

0.30 0.61 0.81 )1.27 0.75 1.44 0.40 )0.86 1.21 1.32 )0.23 )0.77 1.04 1.05 0.49 )0.38

RVI: Dec Mor: Dec Myc: Inc LT50: Inc RVI: Dec Mor: Dec Myc: Dec LT50: Inc RVI: Dec Mor: LevS Myc: Inc LT50: Lev RVI: lev Mor: LevS Myc: Dec LT50: Dec

* Bioassay at a lower concentration may reveal an Allee effect in these isolates. a Isolates with similar response pattern are clubbed. The values represent the actual values of the first isolate in the group. b Conidia/ml: 100 ll/ insect. c Values in brackets indicate fiducial limits, e= error, LT50 could not be computed because the mortality caused did not reach 50%. e Relative virulence index of an isolate at the four doses computed from mortality, mycosis and LT50 using principal component analysis, a data reduction statistical method. The RVI reflects the infection rate of the isolate. f Response pattern over the doses: Inc–Increasing with increase in conidial dose; Dec– Increasing with dose up to a concentration of 107 conidia/ml and decreasing at 108 conidia/ml; Lev–Increasing with dose up to 107 conidia/ml with no further increase at 108 conidia/ml; Levs–Increasing with dose and reaching the theoretical maximum (100%) at a concentration (usually) of 107 conidia/ml with no scope for increase with increase in conidial dose For LT50, values with overlapping fiducial limits are taken as not significantly different.

BB4

BB2

ITCC 4644

ITCC 913

391

392 Dose–infection rate relationship The overall infection rate was assessed from the relative virulence index of an isolate at the four doses tested. In 14 isolates, dose–infection rate showed a linear relationship through all the doses tested, it decreased beyond a dose of 107 conidia/ml in 18 isolates and leveled off at the two high doses in two isolates (Tables 2 and 3). In cases where infectivity increased through all the doses tested, when the mortality leveled off at the two high doses, the increase was either due to an increase in the speed of kill or increase in mycosis or both. Discussion Virulence of a pathogen or parasite is normally dose-dependant with increased virulence correlated with increased doses of the pathogen. In the present study, in a substantial number (19/34) of B. bassiana isolates, the linear relationship between dose and virulence was found to be restricted to only the medium doses. In these isolates, at low doses, no mortality was caused and, with increasing dose from medium to high, no further increase in mortality was observed. In some of these isolates the theoretical maximum of 100% mortality was induced at medium dose and, therefore, no further increase was possible at higher dose; in others, % mortality leveled off from the medium dose even when 100% mortality was not reached. The fact that a low conidial dose was ineffective in inducing mortality with these B. bassiana isolates indicates that a threshold load of infective propagules is required to be effective and dose–mortality relationship is sigmoid. The B. bassiana isolates in which a sigmoid dose–mortality relationship was not detected may be highly virulent isolates in which the threshold value for effective infection is low. Bioassay with a lower conidial dose may reveal sigmoid nature of dose–mortality response in these isolates. It has been noted [4] that an Allee effect is subtle and may be detected only when sufficient range of conidial doses is tested. Bioassays with several different doses spanning a wider range of conidial concentrations in the highly virulent B. bassiana isolates would bring out the sigmoid nature of dose–mortality relationship characteristic of such infection dynamics.

A threshold dose to cause infection may be required due to any of the three mechanisms: (a) the probability of mortality per conidium is constant and independent of the number of conidia applied, but this probability is extremely low so that large numbers of conidia are needed to get measurable infection, or (b) the likelihood of mortality per conidium is highly variable with only a few conidia being actually infective but again this probability is independent of the number of conidia applied or (c) the possibility of mortality per conidium increases as more conidia are applied because the conidia overwhelm host defenses. Only the last of these could really be labeled an Allee effect. It might be possible to distinguish among these mechanisms with additional definitive experiments analyzing the relationship between mean and variance in mortality and number of conidia applied. The results of the present study point to a sigmoid dose–mortality response with several isolates of B. bassiana on M. pustulata. Such a response has also been reported in Metarhizium anisopliae, an entomogenous fungus closely related to B. bassiana with a similar life history pattern [4, 19, 20] and some parasites [21–23]. The basis for the existence of Allee effect has been discussed in Metarhizium [4]. The entry of the infection peg from the conidium into the insect host triggers its second line of defenses – the cellular and humoral immune responses [24]. Successful infection may be possible only when the immune reaction is counteracted and completely saturated. A critical number of conidia may be required to achieve this task. When invaded by fewer conidia, the insect immune system may be successful in overtaking them either through phagocytosis or melanization or encapsulation responses. No infection is thus apparent when the conidial dose is lower than this threshold number. The actual threshold dose for successful infection may depend on the virulence of the fungal isolate. With a highly virulent isolate the threshold value may be low. A highly virulent isolate can more vigorously grow in the host and deplete its nutrients faster. A negative effect of the highest dose tested (108 conidia/ml) on mortality rates was observed with three B. bassiana isolates in beetle bioassays. Goettel et al. [25] reported a similar negative correlation between dose and mortality at concentrations higher than 104 ascospores of Ascosphaera

393 aggregata on larvae of the leaf cutting bee, Megachile rotunda. A negative correlation between dose and the reproductive rate of a parasite has been forecast from theoretical considerations [5]. Such a pattern was observed in lesser mealworm infected with B. bassiana [26]. At the highest doses tested many mealworm larvae that succumbed showed no or few overt symptoms of mycosis [26]. In the present study, such a negative effect of high conidial dose on mycosis was observed with eight isolates. This negative effect of dose on mycosis is reported to be due to premature death of the insect due to toxicosis without ample opportunity for the fungus to grow [27] so that it is rapidly out competed by saprobic antagonists on the insect cadaver [28], or due to scramble competition between the numerous parasites (hyphal bodies) when high doses are applied [29]. In the present study, at the conidial concentrations tested, with 16 B. bassiana isolates, there was a linear increase in the proportion of insects that showed mycosis with increase in dose though mortality levels flattened off from the lower conidial dose. In organisms with semelparous life history, the number of progeny released is a direct reflection of the number of progeny produced at the time of death of the host [4]. In the present bioassays, no count of conidia produced on insect cadavers was made. Only the occurrence or absence of mycosis was scored. The mere occurrence of mycosis does not indicate the number of progeny produced, but it points to the efficiency of the invaded fungal mycelia in the insect host to cross the cuticular barrier effectively and grow and sporulate on its cadaver. Despite scramble competition, when higher doses of parasites are applied, the number of pathogen progeny produced in the host insect at the time of death may be greater than when lower pathogen doses are applied. A larger number may produce more of the enzymes (e.g., chitinase) facilitating emergence from the cuticle after the death of the insect host and external growth of the fungus on the cadaver. Regoes et al. [5] predicted that the life span of the infected hosts should be negatively correlated with dose. Such a trend was observed with seven of the 34 isolates tested here. In a majority (26) of B. bassiana isolates a negative correlation was observed up to a dose of 107 conidia/ml. Above this dose, the life span of the infected host increased (LT50 increased) over the lower concen-

tration (19 isolates) or remained similar (seven isolates). High-dose treatments were found to slow down the growth of B. bassiana isolates [30]. The intra (between individuals of the same isolate) scramble competition intensified at high dose must have contributed to the slower growth resulting in slower death of the insect. From these bioassay studies of B. bassiana on Mylabris pustulata, an Allee effect with a sigmoid relationship between dose and mortality was evident explaining the need for applying high conidial dose to achieve effective results in insect management.

Acknowledgements We are thankful to Dr. R.A. Humber (ARSEF culture collection: Ithaca, New York); Dr. Kerry O’ Donnell (NRRL culture collection; Peoria, Illinois) and the Indian Type Culture Collection (ITCC, IARI, New Delhi) for providing the cultures of B. bassiana. We are thankful to DST, New Delhi, Project No SP/SO/A-10/97, for financial support. CUM is thankful to CSIR for a research fellowship.

References 1. Bartlett MC, Jaronski ST. Mass Production of entomogenous fungi for biological control of insects. In: Burge MN, ed. Fungi in Biological Control Systems, Manchester University Press, Manchester, UK, 1988: 61–85. 2. Wraight SP, Jackson MA, de Kock SL. Production, stabilization and formulation of fungal biocontrol agents. In: Butt TM, Jackson C, Magan N, eds. Fungi as Biocontrol Agents: Progress. Problems and Potential, CAB International, Wallingford, UK, 2001: 253–287. 3. Wraight SP, Caruthers RI. Production, delivery and use of mycoinsecticides for control of insect pests of field crops. In: Hall FR, Menn JJ, eds. Methods in Biotechnology, Vol. 5, Biopesticides: Use and Delivery, Humana Press, Totowa, NJ, 1999: 233–269. 4. Hughes WOH, Petersen KS, Ugelvig LV, Pedersen D, Thomsen L, Poulsen M, Boomsma JJ. Density-dependence and within-host competition in a semelparous parasite of leaf-cutting ants. Evol Biol 2004; 4: 45. 5. Regoes R, Ebert D, Bonhoeffer S. Dose-dependent infection rates of parasites produce the Allee effect in epidemiology. Proc R Soc Lond B 2002; 269: 271–279. 6. Uma Devi K, Sridevi V, Murali Mohan C, Padmavathi J. Effect of high temperature and water stress on in-vitro germination and growth in isolates of the entomopathogenic fungus Beauveria bassiana (Bals.) Vuillemin. J Invert Pathol 2005; 88: 181–188.

394 7.

8.

9.

10.

11.

12.

13.

14. 15.

16. 17. 18. 19.

Butters JA, Uma Devi K, Murali Mohan C, Sridevi V. Screening for tolerance to bavistin, a benzimidazole fungicide containing benzimidazol-2-yl carbamate (MBC), in Beauveria bassiana: Sequence analysis of the b tubulin gene to identify mutations conferring tolerance. Mycol Res 2003; 107: 260–266. Padmavathi J, Uma Devi K, Uma Maheshwara Rao C. The optimum and tolerance pH range is correlated to colonial morphology in isolates of the entomopathogenic fungus Beauveria bassiana – a potential biopesticide. World J. Microbiol Biotechnol 2003; 19: 469–477. Padmavathi J, Uma Devi K, Uma Maheswara Rao C, Nageswara Rao Reddy N. Telomere fingerprinting for assessing chromosome number, isolate typing and recombination in the entomopathogen, Beauveria bassiana (Balsamo) Vuillemin. Mycol Res 2003; 107: 572–580. Uma Devi K, Padmavathi J, Sharma H C, Sitarama N. Evaluation of the virulence of 20 strains of the entomopathogenic fungus Beauveria bassiana (Bals.)Vuillemin to the sorghum shoot borer Chilo partellus Swinhoe (Lepidoptera: Pyralidae) in a laboratory bioassay: their characterization by RAPD-PCR. World J Microbiol Biotechnol 2001; 17: 131–137. Varela A, Morales E. Characterization of some Beauveria bassiana isolates and their virulence toward the coffee berry borer Hypothenemus hampei. J Invertebr Pathol 1996; 67: 147–152. Butt TM, Goettel MS. Bioassays of entomogenous fungi. In: Navon A, Ascher KRS, eds. Bioassays of Entomopathogenic Microbes and Nematodes, CAB International, Wallingford, UK, 2000: 141–195. Goettel MS, Inglis GD. Fungi: Hyphomycetes. In: Lacey LA, ed. Manual of Techniques in Insect Pathology, Academic Press, London, 1997: 213–249. Abbott WS. A method of computing the effectiveness of an insecticide. J Econ Entomol 1925; 18: 265–267. Gomez AK, Gomez AA. Statistical Procedures for Agricultural Research. Singapore: John Wiley & Sons, Inc, 1984. Robertson JL, Preisler HK. Pesticide Bioassay with Arthropods. Boca Raton, Florida: CRC Press, 1992. SPSS Inc. SPSS Standard Version Copyright (c), SPSS for windows Release 7.5.1 (1989–1996), 1996. StatSoft. Stastistica for Windows (Computer Programme Manual) Tulsa, OK: StatSoft, Inc., 1995. Milner RJ, Prior C. Susceptibility of the Australian plague locust, Chortoicetes terminifera, and the wingless grasshopper, Phaulacridium vittatum, to the fungi Metarhizium spp. Biol Control 1994; 4: 132–137.

20. Vestergaard S, Gillespie AT, Butt TM, Schreitter G, Eilenberg J. Pathogenicity of the hyphomycete fungi Verticillium lecanii and Metarhizium anisopliae to the western flower thrips, Frankliniella occidentalis. Biocontrol Sci Technol 1995; 5: 185–192. 21. Agnew P, Koella JC. Life-history interactions with environmental conditions in a host–parasite relationship and the parasite’s mode of transmission. Evol Ecol 1999; 3: 67–89. 22. Little TJ, Ebert D. The cause of parasitic infection in natural populations of Daphnia (Crustacea: Cladocera): The role of host genetics. Proc R Soc Lond B 2000; 267: 2037–2042. 23. Mc Lean AR, Bostock CJ. Scrapie infections initiated at varying doses: an analysis of 117 titration experiments. Philos Trans R Soc Lond B 2000; 355: 1043–1050. 24. Gillespie JP, Bailey AM, Cobb B, Vilcinskas A. Fungi as elicitors of insect immune responses. Arch Insect Biochem Physiol 2000; 44: 49–68. 25. Goettel MS, Vandenberg JD, Duke GM, Schaalje GB. Susceptibility to chalkbrood of alfalfa leafcutter bees, Megachile rotundata, reared on natural and artificial provisions. J Invertebr Pathol 1993; 64: 71–73. 26. Steinkraus DC, Geden CJ, Rutz DA. Susceptibility of lesser mealworm (Coleoptera: Tenebrionidae) to Beauveria bassiana (Moniliales: Moniliaceae): Effects of host stage, substrate, formulation, and host passage. J Med Entomol 1991; 28: 314–321. 27. Glare TR, Milner RJ. Ecology of entomopathogenic fungi. In: Arora DK, Mukerji KG, Drouhet E, eds. Handbook of Applied Mycology, Vol. 2: Humans, Animals and Insects, Marcel Dekker, New York, 1991: 547–612. 28. Fargues J, Remaudie`re G. Considerations of the specificity of entomopathogenic fungi. Mycopathologia 1977; 62: 31–37. 29. Nowak MA, May RM. Superinfection and the evolution of virulence. Proc R Soc Lond B 1994; 255: 81–89. 30. Luz C, Tigano MS, Silva IG, Corderio CM, Aijanabi SM. Selection of Beauveria bassiana and Metarhizium anisopliae isolates to control Triatoma infestans. Memorias do Instituto Oswaldo Cruz 1998; 93: 839–846.

Address for correspondence: K. Uma Devi, Department of Botany, Andhra University, Visakhapatnam, 530 003 AP, India Phone: +91-891-2525582, +91-891-2844563; Fax: +91-8912755547 E-mail: [email protected]